JPS62256075A - Dictionary retrieving system - Google Patents

Abstract

PURPOSE:To properly retrieve a dictionary even when various fluctuations in the notation in an item according to an operator by forming the plural patterns in the fluctuation of the notation of a KANJI (Chinese character) KANA (Japanese syllabary) mixed word and retrieving the dictionary as a character sequence candidate. CONSTITUTION:In a dictionary reading part 5, initially, the 'kakitori' having ka of KANJI, ki of KANA, to of KANJI and ri of KANA of the character code sequence candidate is taken out from a buffer memory 4 and the character code sequence is retrieved from the heading of the dictionary 6. Since there is no heading of the 'kakitori' in the dictionary 6, the second character code sequence candidate 'kakito' having ka of KANJI, ki of KANA and to of KANJI is taken out from the buffer memory 4 and said character code string is retrieved from the heading of the dictionary 6. Since there is no heading of the 'kakito' in the dictionary 6, the third character code sequence candidate 'kakitori' having kaki of KANJI, to of KANJI and ri of KANA is taken out from the buffer memory 4 and said character code sequence is retrieved from the heading of the dictionary 6. Since there is the character code string 'kakitori' kaki or KANJI and to of KANJI and ri of KANA in the item of the dictionary 6, the dictionary reading part 5 reads information corresponding to the item of the 'kakitori' from the dictionary 6.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a speech synthesis device that converts kanji-kana mixed words and sentences into synthesized speech, an information retrieval device that reads information by inputting one item, etc. Concerning dictionary search methods.

[Conventional technology]

In the conventional method, as described in Japanese Patent Application Laid-Open No. 60-24632, input of a kana character string item is converted to a kana code, and input of a kanji character string item is converted to a kanji code and a kana code. The method was to perform a dictionary search.

However, dictionary entries do not only include only kana character strings or only kanji character strings, but also include character strings that include a mixture of kanji and kana, and no consideration was given to such variations in the notation of character strings that include kanji and kana.

[Problem that the invention seeks to solve]

The above-mentioned conventional method does not take into account the dictionary search of character strings containing kanji and kana, and for example, for the dictionary tool 8 "dictori j", it is "dictori".

There was a problem in that when a character string such as ``dictori'' was entered, dictionary searches were impossible. Furthermore, if all possible notations are registered in dictionary items due to such variations in notation, there is a problem in that the capacity of the dictionary becomes enormous.

SUMMARY OF THE INVENTION An object of the present invention is to provide a method that allows a valid dictionary search to be performed on input of a character string containing kanji, kana, and kanji with variations in spelling, without increasing the number of dictionary items.

[Means for solving problems]

The above purpose is to identify whether each character code constituting a dictionary entry or hexagonal entry character string is a kanji code or a kana code, and to leave all kanji codes and preserve the order of the one-character code string. This is achieved by removing one or more codes, generating a combination of character codes, and performing a dictionary search.

[Effect]

In the present invention, a plurality of character strings are generated by leaving only the kanji characters in the kanji-kana-mixed character string and removing the kana characters. When we usually write or write words containing kanji and kana, we sometimes omit the kana characters, for example, ``do'' is used as ``do'', and ``dicti'' is written as ``dictori'' or ``dictori''. However, kanji are not omitted, so it works effectively.

〔Example〕

Hereinafter, one embodiment of the present invention will be described using the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. A character code identification unit 1 is a processing block that identifies whether each character code constituting an input character code string is a kanji code or a kana code, and a character string generation unit 2 is a buffer memory 3 that stores input character strings. A processing block that reads out character codes, generates character string candidates for dictionary search, and stores them in the buffer memory 4. The dictionary reading unit 5 reads character strings generated in the character string generation unit 2 and stores them in the buffer memory 4. This is a processing block that compares the heading of the dictionary 96 and reads dictionary item information if there is a match.

Next, as an example to explain the operation of each processing block using an example, there is a heading "dictori" in the dictionary 6,
Consider the case where the input code string is "dictation".

In the input code character string, the character code identification unit 1 identifies whether each constituent character code is a kanji code or a kana code, and sets the character type flag to 1 if it is a kanji code and 0 if it is a kana code. write to buffer memory 3 along with the character code as ``write''
For input character strings, "calligraphy" and "calligraphy" as shown in Figure 2.
The character type flag for "tori" becomes 1.

"ki" and "su" are 0. However, if the kanji code and kana code can be easily identified from specific bits in the code string or the structure of the code string, the character type flag is not necessary. Among the character codes written in memory 3, the character codes "Ki" and "Tori" whose character type flag is 1 must be left out, and the character codes "Ki" whose character type flag is 0 are retained while preserving the order of the character codes. '' and ``su'', zero or more character codes are removed and stored in a buffer memory 4 that creates a combination of character code strings. In the example of ``dictation'', character code string candidates of ``dictation'', ``dictation'', ``dictation'', and ``dictation'' are stored in the memory buffer 4, as shown in FIG.

The dictionary reading unit 5 first takes out the character code string candidate "Kitori" from the buffer memory 4 and searches for the character code string among the headings in the dictionary 6. Dictionary 6 contains “
Since there is no heading ``Kitori'', next select the second character code string candidate ``Kitori'' from buffer memory 4.
is retrieved and the character code string is searched from among the headings in the dictionary 6.

Since there is no heading ``r dictori'' in the dictionary 6, the third character string code candidate ``dictori'' is retrieved from the buffer memory 4 and searched for the character code string among the headings in the dictionary 6. Since the one-character code string "Kidori" exists in the dictionary 6, the dictionary reading unit 5 reads the dictionary 6.
``Read out the information corresponding to the item of dictation J.''

The above is an embodiment of the present invention, in which a search character string candidate is generated from an input character code string and a dictionary item is searched. In other words, first, the input character code string is searched for a dictionary entry that starts with the same kanji code as the first character of the character code string. Next, the characters of the dictionary entry are searched for. Regarding the code string, identify whether each character code is a kanji code or a kana code, leave all the kanji codes in the character code string, and remove 0 or more kana codes while maintaining the order of the character code string. By doing this, multiple character code string candidates are generated. Then, each of the plurality of character code string candidates is compared with the input character code, and if it matches any one of the plurality of character code string candidates.

It is also possible to configure one or more reading dictionary information corresponding to the dictionary item.

〔Effect of the invention〕

According to the present invention, a plurality of spelling variations for words containing kanji and kana are generated and a dictionary search is performed as a character string candidate, so even if one person has various spelling variations for an item, the dictionary search can be performed correctly. be able to. Furthermore, since there is no need to register a plurality of spelling fluctuation patterns in the dictionary, the dictionary capacity does not need to be increased.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 is a diagram showing an example of a case where the character string "write" is stored in the buffer memory 3 in FIG. 1, and FIG. 3 is a diagram similar to the one shown in FIG. FIG. 4 is a diagram showing an example in which a character string candidate generated from the character string "dicti" is stored in the buffer memory 4 in the computer. 1...Character code identification section, 2...Character string generation section, 3
...Memory buffer, 4...Memory buffer, 5.
...Dictionary reading section, 6...Dictionary, 7...Character code, 8...Character ゛ζ'1, ~

Claims (1)

[Claims] 1. In a dictionary search device that searches a dictionary for an item corresponding to a kanji/kana mixed character code string by notation and reads out the dictionary item information, each character code constituting the dictionary item character code string or input item character code string is a kanji code. Dictionary item string or input item string by identifying whether it is a character code string or a kana code, leaving all kanji codes in the character code string, and removing all kana codes while maintaining the order of the character code string. A dictionary search method characterized by generating a dictionary and performing a dictionary search.